**Association of Salivary Human Papillomavirus Infection and Oral and Oropharyngeal Cancer: A Meta-Analysis**

**Óscar Rapado-González 1,2,3, Cristina Martínez-Reglero 4, Ángel Salgado-Barreira 4, Almudena Rodríguez-Fernández 5, Santiago Aguín-Losada 6, Luis León-Mateos 6, Laura Muinelo-Romay 2,3, Rafael López-López 3,6,\* and María Mercedes Suarez-Cunqueiro 1,3,7,\***


Received: 31 March 2020; Accepted: 26 April 2020; Published: 29 April 2020

**Abstract:** Background. Human papillomavirus (HPV) infection has been recognized as an important risk factor in cancer. The purpose of this systematic review and meta-analysis was to determine the prevalence and effect size of association between salivary HPV DNA and the risk of developing oral and oropharyngeal cancer. Methods. A systematic literature search of PubMed, EMBASE, Web of Science, LILACS, Scopus and the Cochrane Library was performed, without language restrictions or specified start date. Pooled data were analyzed by calculating odds ratios (ORs) and 95% confidence intervals (CIs). Quality assessment was performed using the Newcastle–Ottawa Scale (NOS). Results. A total of 1672 studies were screened and 14 met inclusion criteria for the meta-analysis. The overall prevalence of salivary HPV DNA for oral and oropharyngeal carcinoma was 43.2%, and the prevalence of salivary HPV16 genotype was 27.5%. Pooled results showed a significant association between salivary HPV and oral and oropharyngeal cancer (OR = 4.94; 2.82−8.67), oral cancer (OR = 2.58; 1.67−3.99) and oropharyngeal cancer (OR = 17.71; 6.42−48.84). Significant associations were also found between salivary HPV16 and oral and oropharyngeal cancer (OR = 10.07; 3.65−27.82), oral cancer (OR = 2.95; 1.23−7.08) and oropharyngeal cancer (OR = 38.50; 22.43−66.07). Conclusions. Our meta-analysis demonstrated the association between salivary HPV infection and the incidence of oral and oropharyngeal cancer indicating its value as a predictive indicator.

**Keywords:** human papillomavirus; oral cancer; oropharyngeal cancer; saliva; meta-analysis

#### **1. Introduction**

Human papillomavirus (HPV) infection has been recognized as an important risk factor in a subset of head and neck squamous cell carcinomas, independently of traditional risk factors such as tobacco or alcohol use [1,2]. Globally, around 38,000 cases of head and neck cancer are attributed to the HPV infection. Of these, around 76% are cases of oropharynx cancer, 12% of oral cavity cancer and 10% of larynx cancer [3]. Currently, it is well known that HPV-status determines the molecular landscape of these tumors and their clinical evolution, with a better prognosis and response to therapy being found in HPV-positive patients [4,5].

HPVs are small, non-enveloped, close-circular, double-stranded DNA viruses of approximately 8000 base-pairs which present a specific tissue tropism infecting epithelial cells of the skin and mucosae of the anogenital and upper aero-digestive tract [6]. More than 200 different HPV types have been identified and classified into low-risk and high-risk according to their oncogenic potential. In this sense, high-risk HPV (HR-HPV) can promote the malignant transformation of HPV-infected cells through E6 and E7 viral oncoproteins, responsible for inactivating the *TP53* and *Rb* (retinoblastoma tumor suppressor gene) [7]. A subset of 12 alpha HR-HPV (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, and 59) has been classified as carcinogenic to humans according to the International Agency of Research in Cancer [8]. HR-HPV is considered the main cause of cervical cancer, genotypes 16 and 18 being responsible for 70% of cases [9]. In addition, several studies have also demonstrated the pathogenic role of HPV in other anogenital cancers [10–12] as well as in head and neck cancers [13]. Currently, HPV16 is widely recognized as an etiological factor in oropharynx tumors [14], however, not enough evidence exists regarding the HPV relationship and the anatomic subsites of head and neck squamous cell carcinoma [15].

Nowadays, a variety of molecular biological methods have been developed for the detection and genotyping of HPV at DNA, mRNA, and protein levels by polymerase chain reaction (PCR), real-time PCR, in situ hybridization, immunohistochemistry and serum antibody assays [16]. In addition, next-generation HPV sequencing approaches provide accurate information on genotype composition and pathways to better understand functional consequences [17]. Certain collection approaches present difficulties. For example, tumoral tissue biopsy is invasive and tumors may be inaccessible. For its part, the collection of oral exfoliated cells with cotton swabs or cytobrush is restricted to a specific and accessible oral area, making collection difficult for non-visual tumors and early molecular alterations. To overcome these drawbacks, the detection of HPV in oral exfoliated cells from saliva (with or without oral rinses) represents a quick and easy non-invasive alternative for oral and oropharyngeal cancer screening in high-risk populations. In this sense, several researchers have analyzed the prevalence of salivary HPV DNA from head and neck cancer, however, to our knowledge, no previous systematic review has elucidated evidence of this relationship. Therefore, the aim of the present systematic review and meta-analysis was to determine the prevalence and effect size of association between salivary HPV DNA and the risk of developing oral and oropharyngeal cancer.

#### **2. Materials and Methods**

#### *2.1. Protocol and Registration*

This study was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [18] and the protocol was registered with the International Prospective Register of Systematic Reviews (reference No. CRD42020161345).

#### *2.2. Search Strategy and Study Selection*

The systematic literature search was performed in PubMed, EMBASE, Web of Science, LILACS, Scopus and the Cochrane Library through 9 January 2020, without language restrictions or specified start date. The following combinations of keywords and medical subject headings were used: (human papilloma virus OR HPV) AND (saliva OR oral rinses OR mouthwash) AND (oral squamous

cell carcinoma OR OSCC OR oropharyngeal squamous cell carcinoma OR OPSCC OR oral cancer OR oropharyngeal cancer). All studies were screened based on the title and abstract, and eligible manuscripts were retrieved for full-text review. Additionally, we manually searched the reference lists in each original and review article in order to avoid missing potential studies. The literature search was performed independently by two researchers (ORG and MMSC), and any disagreements were resolved by consensus. The studies selected through the search strategy and other references were managed using RefWorks software, and duplicated items were removed using the associated tools.

#### *2.3. Eligibility Criteria*

We included the studies that met the following criteria: (1) case-control studies of patients with oral and/or oropharyngeal cancer and healthy controls, (2) HPV DNA prevalence determined in salivary samples (whole saliva or oral rinses), and (3) sufficient data to calculate odds ratios (ORs) with 95% confidence intervals (CIs). The exclusion criteria were as follows: (1) in vitro or animal study, (2) reviews, letters, personal opinions, book chapters, case reports, and conference abstracts, and (3) duplicate articles or suspicion of data overlap.

#### *2.4. Protocol and Registration*

Two researchers (ORG and MMSC) independently assessed each eligible manuscript, extracted data using a pre-established form, and collated the data into a Microsoft Excel spreadsheet (Microsoft Corp. Redmond, WA, USA). Any disagreement among reviewers was resolved by consensus. The following information was extracted from each study: author, publication year, country, type of sample, method of collection, tumor location, sample size, HPV detection method, number of cases and HPV-positive cases, number of controls and HPV-positive controls, HPV-positive genotypes, overall HPV DNA prevalence (number of subjects testing positive for any HPV type) and type-specific HPV DNA prevalence (number of subjects testing positive for specific HPV types: HPV16 or HPV18, HR-HPV and LR-HPV). If the required data were incomplete, attempts were made to contact the authors to obtain the missing information.

#### *2.5. Assessment of Risk Bias*

The Newcastle-Ottawa Scale (NOS) [19] was used to evaluate the individual quality of the selected studies by three independent researchers (ORG, ARF, and MMSC), and discrepancies were resolved by consensus. The NOS assesses the quality of non-randomized studies based on design, content and ease of use directed to the task of incorporating the quality assessments in the interpretation of meta-analytic results. This 'star system' consists of 8 items classified into three broad perspectives: the selection of study groups; the comparability of the groups; and the ascertainment of either the exposure or outcome of interest for case-control or cohort studies. The highest quality studies were allotted a maximum of one star for each item, except for, the item related to comparability, which was allowed the assignment of a maximum of two stars. The NOS score ranged from 0 to 9 stars and validity criteria were as follows: 8–9, high quality; 6–7, medium quality; <5 low quality.

#### *2.6. Statistical Analysis*

Statistical analysis was conducted using the meta package of free R software (v.3.6.2; https: //www.r-project.org). Firstly, to evaluate the statistical model applied to the meta analytic database, heterogeneity was assessed using the Cochran's Q statistic test-based Chi-squared test and I2 statistics. Heterogeneity was considered significant when I2 > 50% and/or presence of a *p* < 0.10 for the Cochran's Q test. The prevalence of HPV DNA and HPV genotypes in oral and/or oropharyngeal cancer was calculated using fixed or random effects depending on the heterogeneity. The relationship between saliva HPV DNA infection and oral and/oropharyngeal cancer risk was evaluated by pooled odds ratio (OR) and 95% confidence intervals (CIs) comparing cases to controls. If significant heterogeneity was detected, the DerSimonian and Laird random-effects model was applied to calculate the pooled OR

with 95% CIs; otherwise, the Mantel–Haenszel fixed-effects model was used. Then, subgroup analyses were performed to explore the potential sources of heterogeneity among studies according to the anatomic tumor location and HPV genotypes. Additionally, publication bias was checked with Begg's and Egger's tests and by visual inspection in funnel plots demonstrating the relationship between the individual log ORs and their standard errors [20,21]. *p*-values of <0.05 were considered to indicate statistical significance.

#### **3. Results**

#### *3.1. Study Selection*

A total of 1669 articles were identified across the six electronic databases and three additional reports from the reference lists. After removing duplicates, a total of 1542 articles were screened based on the title and abstract, and 1494 were excluded for lack of adherence to our inclusion criteria. Therefore, full-text articles were retrieved for the remaining 48 articles. After a full-text review, 34 articles were excluded for the following reasons: non case-control studies (22); controls under risk conditions (2); suspicious of data overlap (3); insufficient data (3); and reviews, letters, and meta-analysis (4). Finally, 14 articles met all the inclusion criteria and were included in the final analysis. A detailed flowchart showing the selection process is shown in Figure 1.

**Figure 1.** Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) flow diagram of the literature selection process, including identification, screening, eligibility and total studies included in qualitative and quantitative synthesis.

#### *3.2. Study Characteristics*

Individual characteristics of the included studies are summarized in Table 1. A total of 14 articles evaluating HPV prevalence in oral and/or oropharyngeal cancer were included in this meta-analysis, and these studies were carried out from 2005 to 2019. Study sample sizes ranged from 42 to 677 subjects. The study units in this meta-analysis comprised a total of 2320 cases (658 from the oral cavity, 1160 from the oral cavity plus oropharynx and 502 from the oropharynx), and 5868 controls (2210 from the oral cavity, 2304 from the oral cavity plus oropharynx and 1354 from the oropharynx). As reported in Table 1, four studies were conducted in India [22–25], three in the USA [26–28], and two in Sweden [29,30], whereas the remaining studies were carried out in the following countries: Canada [31], France [32], Hungary [33], Pakistan [34], and Iran [35]. In terms of sampling, oral rinses and saliva (*n* = 7, 50%, respectively) were analyzed for HPV positivity and genotyping. The methods most used for saliva HPV-DNA determination were conventional PCR, nested PCR and quantitative PCR. However, other analytical strategies such as next generation sequencing or immunoassays were also employed for salivary HPV genotyping (Table 1).

#### *3.3. Study Quality*

Assessment of risk of bias and quality was performed according to NOS (Table S1). Regarding the selection domain, adequate description about characteristics and selection criteria for cases and controls were provided by all of the included studies. Regarding the comparability domain, six out of the 14 studies matched for age and at least one additional factor. Insofar as the exposure domain, few studies reported the blinding of analyses or non-response rates. The mean NOS score in our meta-analysis was six.

#### *3.4. Meta-Analysis*

#### 3.4.1. Salivary HPV Association with Oral and Oropharyngeal Cancer

Overall, the prevalence of salivary HPV for oral and oropharyngeal carcinoma was of 43.2% (*n* = 1160) while the infection rate in the healthy control group was of 8.9% (*n* = 2304). Salivary HPV16 was the most common type of HPV DNA positive cases (*n* = 1116), representing 27.5% (Figure 2).

**Figure 2.** Schematic drawing of salivary HPV and prevalence of oral and/or oropharyngeal cancer. Oral tissue sheds pathogen-infected cells containing different HPV DNA genotypes (HPV16, HPV18, HR-HPV, and LR-HPV) into saliva (with or without oral rinses). The prevalence of salivary HPV DNA varied according to anatomic tumor location, showing the highest infection rate in oropharyngeal carcinomas. In addition, the type-specific prevalence in saliva was also different according to the anatomic tumor location.


*J. Clin. Med.* **2020**, *9*, 1305



54, 56, 59, 61, 62, 66, 68, 69, 70, 71, 72, 73, 81, 82, 83, 84, and 89.

#### *J. Clin. Med.* **2020**, *9*, 1305

Our meta-analysis included a total of 1160 cases and 2304 controls. The pooled analysis showed a significant association between positive salivary HPV DNA status and oral and oropharyngeal cancer with a pooled OR of 4.94 (95% CI = 2.82−8.67; *p* < 0.01) (Figure 3).


**Figure 3.** Forest plot for the studies on the association between salivary HPV and oral and oropharyngeal cancer. The squares indicate the ORs (odds ratios) in each study, with square sizes inversely proportional to the standard error of the OR. The diamond shape indicates the pooled ORs. Horizontal lines represent 95% CIs (confidence intervals), I2 > 50% indicates severe heterogeneity.

A random-effects model was used because heterogeneity was identified among the 14 studies (I2 = 82%). Visual inspection of the funnel plot revealed a symmetrical (Egger's test, *p* = 0.159; Begg's test, *p* = 0.298) distribution of the studies, indicating no evidence of publication bias (Figure 4).

**Figure 4.** Funnel plot for studies (of 14 studies) on the association between salivary HPV and oral and oropharyngeal cancer. The vertical line represents the pooled OR using random-effect meta-analysis. Two diagonal lines represent (pseudo) 95% confidence limits around the OR for each standard error on the vertical axis. In the absence of heterogeneity, 95% of the studies should lie within the funnel defined by these diagonal lines. Abbreviations: se OR, standard error of odds ratio.

For the type-specific analysis (Figure 5), salivary HPV16 showed a significant association with a pooled OR of 10.07 (95% CI = 3.65−27.82; *p* < 0.01). However, salivary HPV18 did not show any significant increased risk for oral and oropharyngeal cancer with a pooled OR of 1.80 (95% CI = 0.66−4.90). In addition, a significant association was found for salivary HR-HPV with OR of 5.94 (95% CI = 2.78−12.69; *p* < 0.01), whereas salivary LR-HPV did not show any significant increased risk with OR of 1.45 (95% CI = 0.70−2.98). The respective funnel plots are represented in Figures S1–S4.


**Figure 5.** Forest plot for the studies on the association between salivary HPV and oral and oropharyngeal cancer. The squares indicate the ORs in each study, with square sizes inversely proportional to the standard error of the OR. The diamond shape indicates the pooled ORs. Horizontal lines represent 95% CIs. I2 > 50% indicates severe heterogeneity. (**a**) HPV16, (**b**) HPV18, (**c**) HR-HPV, and (**d**) LR-HPV.

#### 3.4.2. Type-Specific Salivary HPV Association with Oropharyngeal Cancer

Our subgroup meta-analysis consisted of eight studies, including 502 cases and 1354 controls. In the pooled analysis, salivary HPV DNA infection and oropharyngeal cancer showed a significant association with a pooled OR of 17.71 (95% CI = 6.42−48.84; *p* < 0.01) (Figure 6).


**Figure 6.** Forest plot for the studies on the association between salivary HPV and anatomic tumor subsites. The squares indicate the ORs in each study, with square sizes inversely proportional to the standard error of the OR. The diamond shape indicates the pooled ORs. Horizontal lines represent 95% CIs. I2 > 50% indicates severe heterogeneity. (**a**) Oral Cancer and (**b**) Oropharyngeal Cancer.

According to type-specific analysis (Figure 7), salivary HPV16 showed a significant association with a pooled OR of 38.50 (95% CI = 22.43−66.07; *p* < 0.01) whereas salivary HPV18 showed no significant association with a pooled OR of 1.92 (95% CI = 0.63−5.91). In addition, a significant association was found for salivary HR-HPV with a pooled OR of 26.69 (95% CI = 3.46−206.17; *p* < 0.01) whereas no significant association was found for salivary LR-HPV with a pooled OR of 2.08 (95% CI = 0.75−5.81). Their respective funnel plots are shown in Figures S5–S9.


**Figure 7.** Forest plot for the studies on the association between salivary HPV and oropharyngeal cancer. The squares indicate the ORs in each study, with square sizes inversely proportional to the standard error of the OR. The diamond shape indicates the pooled ORs. Horizontal lines represent 95% CIs. I2 > 50% indicates severe heterogeneity. (**a**) HPV16, (**b**) HPV18, (**c**) HR-HPV, and (**d**) LR-HPV.

#### 3.4.3. Type-Specific Salivary HPV Association with Oral Cancer

Our subgroup meta-analysis consisted of 12 studies, including 658 cases and 2210 controls. In the pooled analysis, salivary HPV DNA infection and oral cancer showed a significant association with a pooled OR of 2.58 (95% CI = 1.67−3.99; *p* < 0.01) (Figure 6). According to type-specific analysis (Figure 8), salivary HPV16 showed a significant association with a pooled OR of 2.95 (95% CI = 1.23−7.08; *p* = 0.02), whereas no significant association was observed for salivary HPV18 with a pooled OR of 1.51 (95% CI = 0.45−5.15). In addition, a significant association was observed for salivary HR-HPV with a pooled OR of 4.44 (95% CI = 2.47−7.98; *p* < 0.01). However, salivary LR-HPV did not show any significantly increased risk for oral cancer with OR of 1.79 (95% CI = 0.67−4.74). Their respective funnel plots are shown in Figures S10–S14.


**Figure 8.** Forest plot for the studies on the association between salivary HPV and oral cancer. The squares indicate the ORs in each study, with square sizes inversely proportional to the standard error of the OR. The diamond shape indicates the pooled ORs. Horizontal lines represent 95% CIs. I2 > 50% indicates severe heterogeneity. (**a**) HPV16, (**b**) HPV18, (**c**) HR-HPV, and (**d**) LR-HPV.

#### **4. Discussion**

In the present study the overall pooled prevalence of salivary HPV-related to oral and oropharyngeal cancer was 43.2%. Similarly, a meta-analysis based on 11 case-control studies evaluating the HPV infection in oral and oropharyngeal cancer found an HPV DNA prevalence of 39.27% [36]. In terms of anatomic tumor location, we observed the highest prevalence of salivary HPV in oropharyngeal cancer (51.9%), whereas the overall percentage in the oral cavity was 32.5%. Similarly, a large comprehensive meta-analysis based on data from 148 studies estimated a pooled HPV DNA prevalence of 45.8% in oropharynx tumors and 24.5% in oral cavity tumors [37]. Although our meta-analysis did not evaluate HPV DNA prevalence in different oropharynx subsites, evidence shows that HPV is most prevalent in tonsils and base of tongue cancers compared to tumors located in walls of oropharynx, uvula and soft palate [38].

Overall, in our study, salivary HPV16 was the most commonly detected oncogenic type, accounting for around 28% of cases. As we expected, salivary HPV16 showed a higher prevalence in oropharyngeal cancer (39.6%) than oral cancer (18.6%), in accordance with previous studies [37,39,40]. In particular, the salivary HPV16 prevalence in our study was slightly higher in oral cancer than reported in a meta-analysis by Nydiae et al. [37] (18.6% vs 14.9%), however, other authors have reported higher rates of HPV16 prevalence in oral carcinoma, ranging from 20% to 50% [41–43]. Salivary HPV18 was another oncogenic HPV type commonly evaluated by the included studies. Unlike salivary HPV16, HPV18 positivity was found much less frequently, with an overall prevalence of 2.3%. Salivary HPV18 prevalence was even lower in oropharynx tumors (1.7%) as compared to oral cavity tumors (2.7%). One plausible explanation for the decreased prevalence of salivary HPV18 in both oral and oropharyngeal cancers is its specific tropism for glandular tissue and adenocarcinomas, while most head and neck cancers are predominantly of the squamous cell carcinoma type [44]. In addition, HR-HPV has developed a variety of mechanisms facilitating HPV evasion of recognition and clearance by the host immune system [45], which probably contributes to the different viral persistence in each of the anatomic regions of the head and neck. As in our study, Kreimer et al. [40] and Ndyae et al. [37] also found a low HPV18 prevalence in oropharynx tumors (1% and 0.7%, respectively), however, these studies reported a higher HPV18 prevalence in oral cancer (8% and 5.9%, respectively). These differences could be explained by the effect on HPV prevalence of different covariates such as geographical location, lifestyles (alcohol, tobacco or sexual activity), sample size, types of samples and methods used for HPV detection.

To the best of our knowledge, this is the first meta-analysis evaluating the association between salivary HPV and oral and/or oropharyngeal cancer. The pooled OR showed that oral and oropharyngeal cancer patients had an almost five-fold higher risk of HPV infection than controls. A previous meta-analysis evaluating the presence of HPV in oral and oropharyngeal cancer detected by different methods (histopathology, serum analysis, and cytopathology using OralCDx or oral swishes) reported a significant association with an OR of 2.82 [36]. Overall, our results indicate that salivary HPV causes a higher risk of oral and oropharyngeal carcinogenesis. In addition, we also conducted different subgroup analysis to evaluate the impact of HPV infection on cancer risk according to anatomic tumor location and HPV genotypes. We stratified the salivary HPV studies by anatomical location observing a stronger association between salivary HPV and oropharynx tumors compared to oral cavity tumors. Similarly, Shaik et al. performed a comprehensive metanalysis of HPV-associated head and neck cancers, reporting the highest association for oropharyngeal cancer, with an OR of 14.66, whereas oral cavity and laryngeal cancers had ORs of 4.06 and 3.23, respectively [46]. In addition, we evaluated type-specific salivary HPV risk associated with oral and oropharyngeal cancer. Compared to oncogenic potential, salivary HR-HPV types were associated with an increased risk of oral and oropharyngeal carcinomas. Thus, salivary HPV16 was significantly associated with oral cancer, confirming the findings reported in previous studies [43,47]. Like our study, Hobbs et al. reported a weak statistically significant association between HPV16 and oral cancer, with an OR of 2.0 [45]. On the contrary, a higher association (OR = 9) was reported by Zhu et al., suggesting the potential oncogenic role of HPV16 in oral carcinogenesis in Chinese population [43]. However, in our study, a stronger association was found between salivary HPV16 and oropharyngeal cancer, presenting an OR of 38.50, which suggests the role of HPV16 in the etiology of oropharyngeal cancer. Unlike other meta-analysis [47,48], our study did not analyze association based on specific subsites of the oropharynx. In this sense, a consistent association between HPV16 infection and tonsil cancer has previously been described [47], which seems to indicate a different oncogenic role for HPV infection in the different subsites of oropharynx.

All the studies included in the present meta-analysis addressed HPV status in oral exfoliated cells collected from saliva with or without oral rinses. In this sense, the first association between oral HPV and oral cancer was reported by Smith et al. [49]. These authors evaluated HPV status in oral exfoliated cells collected by oral rinses from 93 patients and 205 controls finding significantly increased risk (OR = 3.70) of cancer in positive oral HPV patients regardless of alcohol and tobacco use [49]. According to the evidence, salivary HPV DNA represents a promising approach for identifying oral HPV infection. Several authors have shown a significant correlation between HPV DNA detected

in tissue and positivity for HPV DNA in saliva, suggesting the potential value of this biofluid for detecting HPV and thus predicting HPV-related head and neck carcinomas [1,50]. Furthermore, salivary HPV DNA has demonstrated to be a good marker for detecting HPV in oropharyngeal cancer, as a high agreement between salivary HPV16 DNA infection and tumor p16 expression has been observed [51–53]. However, a recent study revealed a lower sensitivity for identifying p16-positive oral cancer patients through salivary HPV, which may indicate a limited involvement of HPV16 in oral carcinogenesis [54]. Interestingly, our study reviewed the different salivary HPV genotypes identified in oral and oropharyngeal carcinomas, providing additional evidence on the co-existence of multiple HPV types during carcinogenesis. In this matter, saliva analysis represents a great opportunity for the identification and characterization of novel HPVs involved in head and neck cancer.

Our study has several strengths. It is the first meta-analysis highlighting the association between salivary HPV infection and oral and/or oropharyngeal cancer. Moreover, we examined both the overall and the specific prevalence of salivary HPV DNA in oral and/or oropharyngeal cancer. In addition, we performed a comprehensive literature review without language restrictions and the results of our study were in concordance with the scientific evidence. However, the present study is not exempt from limitations. Firstly, the studies included in our meta-analysis were heterogeneous, which could be explained by different factors such as ethnicity, sample size, geographic region, anatomic tumor location, method of HPV detection and different HPV genotypes. Although we performed a subgroup analysis by anatomic tumor location and HPV genotypes, we were unable to elucidate the potential sources contributing to this heterogeneity. Secondly, data such as age, smoking, drinking, sexual habits or diet were not provided by the studies in our sample, hampering the assessment of these confounding variables. Thirdly, some studies included in our analysis could be biased due to the fact that cases and controls were not matched for demographic variables such as age, sex and lifestyle habits. In addition, although almost all these studies analyzed HPV16 and HPV18, we observed high variability regarding HPV genotypes and HPV detection methods, which could substantially affect the results of our analysis.

#### **5. Conclusions**

To the best of our knowledge, this is the first meta-analysis addressing the association between salivary HPV infection and oral and oropharyngeal carcinoma. The findings of this meta-analysis provide additional evidence that salivary HPV is associated with oral and oropharyngeal cancer, suggesting that salivary HPV infection is a risk factor for oral and oropharyngeal cancer. However, to validate our findings, future research should focus on prospective cohort studies that explore the occurrence of salivary HPV infection in oral and oropharyngeal cancer. In addition, it is necessary to analyze confounding variables that might be associated with an increased risk of HPV infection in oral and oropharyngeal cancer.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2077-0383/9/5/1305/s1, Table S1: The Newcastle–Ottawa Scale (NOS) for assessing the quality of included studies, Figure S1: funnel plot for studies (of 12 studies) on the association between salivary HPV16 and oral and oropharyngeal cancer, Figure S2: funnel plot for studies (of 10 studies) on the association between salivary HPV18 and oral and oropharyngeal cancer, Figure S3: funnel plot for studies (of 12 studies) on the association between salivary HR-HPV and oral and oropharyngeal cancer, Figure S4: funnel plot for studies (of five studies) on the association between salivary LR-HPV and oral and oropharyngeal cancer, Figure S5: funnel plot for studies (of eight studies) on the association between salivary HPV and oropharyngeal cancer, Figure S6: funnel plot for studies (of seven studies) on the association between salivary HPV16 and oropharyngeal cancer, Figure S7: funnel plot for studies (of five studies) on the association between salivary HPV18 and oropharyngeal cancer, Figure S8: funnel plot for studies (of five studies) on the association between salivary HR-HPV and oropharyngeal cancer, Figure S9: funnel plot for studies (of three studies) on the association between salivary LR-HPV and oropharyngeal cancer, Figure S10: funnel plot for studies (of 12 studies) on the association between salivary HPV and oral cancer, Figure S11: funnel plot for studies (of eight studies) on the association between salivary HPV16 and oral cancer, Figure S12: funnel plot for studies (of seven studies) on the association between salivary HPV18 and oral cancer, Figure S13: funnel plot for studies (of nine studies) on the association between salivary HR-HPV and oral cancer, and Figure S14: funnel plot for studies (of two studies) on the association between salivary LR-HPV and oral cancer.

**Author Contributions:** Conceptualization, Ó.R.-G., R.L.-L. and M.M.S.-C.; methodology, Ó.R.-G., M.M.S.-C., Á.S.-B., and C.M.-R.; software, Á.S.-B. and C.M.-R.; validation, Ó.R.-G. and M.M.S.-C.; formal analysis, Á.S.-B. and C.M.-R.; investigation, Ó.R.-G. and M.M.S.-C.; resources, Ó.R.-G. and A.R.-F.; data curation, Á.S.-B. and C.M.-R.; writing—original draft preparation, Ó.R.-G. and M.M.S.-C.; writing—review and editing, Ó.R.-G., M.M.S.-C., L.M.-R., R.L.-L., S.A.-L. and L.L.-M.; visualization, Ó.R.-G. and M.M.S.-C.; supervision, M.M.S.-C. and R.L.-L.; project administration, M.M.S.-C. and R.L.-L.; funding acquisition, M.M.S.-C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** R.L.-L. reports other from Nasasbiotech, during the conduct of the study; grants and personal fees from Roche, grants and personal fees from Merck, personal fees from AstraZeneca, personal fees from Bayer, personal fees and non-financial support from BMS, personal fees from Pharmamar, personal fees from Leo, outside the submitted work. The rest of the authors have nothing to disclose.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Review* **Saliva and Oral Diseases**

#### **Emanuela Martina** †**, Anna Campanati \*,**†**, Federico Diotallevi and Annamaria O**ffi**dani**

Dermatology Clinic, Department of Clinical and Molecular Sciences, Polytechnic Marche University, via Conca 71, 60126 Ancona, Italy; ema.martina@gmail.com (E.M.); federico.diotallevi@hotmail.it (F.D.); a.offidani@ospedaliriuniti.marche.it (A.O.)

**\*** Correspondence: anna.campanati@gmail.com; Tel.: +39-0715963433

† These authors contributed equally to the manuscript.

Received: 29 December 2019; Accepted: 3 February 2020; Published: 8 February 2020

**Abstract:** Saliva is a fascinating biological fluid which has all the features of a perfect diagnostic tool. In fact, its collection is rapid, simple, and noninvasive. Thanks to several transport mechanisms and its intimate contact with crevicular fluid, saliva contains hundreds of proteins deriving from plasma. Advances in analytical techniques have opened a new era—called "salivaomics"—that investigates the salivary proteome, transcriptome, microRNAs, metabolome, and microbiome. In recent years, researchers have tried to find salivary biomarkers for oral and systemic diseases with various protocols and technologies. The review aspires to provide an overall perspective of salivary biomarkers concerning oral diseases such as lichen planus, oral cancer, blistering diseases, and psoriasis. Saliva has proved to be a promising substrate for the early detection of oral diseases and the evaluation of therapeutic response. However, the wide variation in sampling, processing, and measuring of salivary elements still represents a limit for the application in clinical practice.

**Keywords:** biomarkers; saliva; oral cancer; oral lichen planus; psoriasis; oral diseases

#### **1. Introduction**

Saliva is a biological fluid secreted by major and minor salivary glands. The major salivary glands are the parotid, submandibular, and sublingual glands. Minor salivary glands are widely disseminated throughout the entire oral cavity. Saliva provides lubrication; facilitates mastication, digestion, and taste; it has antimicrobial properties; and serves as buffer for acidic food. Moreover, saliva inhibits the demineralization of teeth and protects from caries [1]. The physiological secretion generates 0.75–1.5 L per day, with a decrease during the night [2]. Saliva contains 99% water and proteins for the remaining 1% (mucins, enzymes, immunoglobulins), electrolytes, lipids, and inorganic substances [3].

There are many advantages to employing saliva as a substrate for diagnostic analysis. Its sampling is fast, inexpensive, non-invasive, and well tolerated by children and people with disabilities; moreover, it is a safe procedure for healthcare providers [4]. Many serum substances enter saliva through passive diffusion, active transport, or extracellular ultrafiltration [5]. Obviously, compared with blood, levels of several analytes are lower, which was an obstacle until a few years ago [6]. Nowadays, highly sensitive molecular methods are available and can be used in the detection of many elements in saliva, despite their dimensions and concentrations [7].

In recent decades, enormous progress has been made in early diagnosis and screening for many diseases, especially for neoplastic conditions. However, some of these methods are invasive or expensive, and for certain conditions, accurate tests are still not available. This is the case for oral cancer, the sixth most common cancer worldwide, frequently diagnosed at an advanced stage with a 5 year survival rate of 50% [8].

In accordance with Biomarkers Definitions Working Group 2011, a biomarker is a characteristic that can be objectively measured and evaluated as indicator of normal biological or pathogenic processes, or as an indicator of pharmacologic response to therapeutic interventions [9]. The detection of salivary biomarkers and their use in clinical practice in the near future is one of the most ambitious aims of contemporary researchers.

#### **2. Materials and Methods**

The review was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) checklist. A search in the PubMed database was carried out using the keywords "saliva", "salivary", "biomarkers", "oral diseases", "oral lichen planus", "oral cancer", and associations between terms. We selected only articles written in English. The papers were selected first by analyzing titles and abstracts, in order to choose a correct match with our topic; full-text articles were then studied and included in the revision.

#### **3. Sampling and Processing Techniques**

Many factors can alter the composition and total amount of saliva. The time of day, hydration, body position, drugs intake, smoking, psychological stimuli, food assumption, and other factors related to systemic conditions can change the characteristics of saliva in a single subject [10]. A sample of saliva can be collected at rest or after stimulation. This procedure consists of offering a gum or swab to chew, or specific taste stimuli such as citric acid [11]. The stimulation changes not only the volume, but also the composition of saliva; it has been demonstrated that parasympathetic stimulation produces a high flow rate, but sympathetic stimulation produces a small flow richer in proteins and peptides [12]. Consequently, proteome profile and proportion are changeable as a reaction to neural activation [13].

As regards clinical trials, saliva is usually collected at rest ("unstimulated saliva") after at least 1 h of fasting, without drinking or smoking; the patient must be comfortably seated, avoid oro-facial movements for 5 min, and, just before the sampling, has to rinse their mouth with deionized water [11].

Saliva specimens can be collected from whole saliva or from a single gland (for example, the parotid gland). This procedure, which uses a different method, can be uncomfortable for patients and therefore is rarely used [14]. It should be specified that whole saliva has a higher proportion of non-salivary materials such as food debris, bacteria, desquamated epithelial cells, and leukocytes [15].

The gold standard method is to drain saliva using special devices (Salivette®, Sarstedt, Nümbrecht, Germany; Quantisal®, Immunalysis, Pomona, CA, USA; Orapette®, Trinity Biotech, Dublin, Ireland and SCS® Greiner Bio-One, Kremsmünster, Austria) [16].

Controversies are evident in the literature regarding centrifugation and speed, addition of PIC (protease inhibitor cocktail), and storage temperature. Most authors recommended the use of a protease inhibitor mixture in order to stabilize the substrate; moreover, the samples collected must be immediately stored in ice containers and, after processing, stored at −80 ◦C [17]. All these steps are necessary for bacterial growth inhibition and the minimal impairment of salivary proteins.

#### **4. "Salivaomics"**

The term "salivaomics" was coined in 2008 to emphasize the various "omics" found in saliva: genome, transcriptome, proteome, metabolome, and microbiome [18]. Salivaomics has been widely studied in recent years thanks to the advent of more advanced analytical techniques. Nearly 70% of the genome in saliva is human; the remaining 30% belongs to the oral microbiota [19]. The DNA contained in saliva is approximately 24 μg (range 0.2–52 μg), which is almost 10 times lower than in blood, but genotyping techniques require as little as 5 ng/mL of DNA to work effectively [20]. Polymerase chain reaction (PCR) and sequencing arrays can be applied to saliva samples. The analysis of salivary DNA aims especially to detect aberrant DNA methylation, which is the first epigenetic mark of neoplastic alterations [21].

#### *4.1. Transcriptomes*

mRNA and microRNA secreted from cells can be easily detected in saliva. Reverse transcriptase polymerase chain reaction and microarray are the most commonly used analyses. Zhang et al. first developed a technique to permit stabilization and to process salivary RNA [22]. The great potential of transcriptome study in the early detection of cancer and other diseases has been reported [23–25]. More recently, noncoding RNAs (ncRNAs) or microRNAs (miRNA) have been the subject of many studies because of their role in oncogenesis and their great stability in biological fluids, including saliva [26]. MicroRNAs are encoded by genes but are not translated into proteins; it is now generally accepted that these small nucleotides are involved in cell differentiation, proliferation, and survival. Moreover, many studies have already demonstrated the dysregulation of miRNAs in cancer tissues [27,28]. Surprisingly, salivary microRNAs are more stable than mRNAs, which makes this biological fluid a suitable substrate for transcriptome analysis.

#### *4.2. Metabolome*

The endogenous metabolites are nucleic acids, vitamins, lipids, organic acids, carbohydrates, thiols, and amino acids. The study of the salivary metabolome can provide an overview of the general health status or modification during systemic diseases [29]. In 2010, Sugimoto et al. first used salivary metabolome analysis with capillary electrophoresis and mass spectrometry to detect differences between healthy controls and patients with solid cancer [30]. The authors identified three metabolites that were oral-cancer-specific and eight metabolites that were pancreatic-cancer-specific. Nuclear magnetic resonance (NMR) spectroscopy can detect and measure metabolites in a solution with minimal sample preparation. This quantitative technique is based on the magnetic properties of atomic nuclei [31]. Each compound has a characteristic resonance frequency that makes it easy to distinguish. Moreover, the area under a signal peak is proportional to the concentration of the metabolite [32]. Liquid chromatography–mass spectrometry (LC-MS) is considered the gold standard in metabolomics. In fact, it is able to analyze an enormous range of analytes with a greater sensitivity than NMR [33]. This technique provides very high chromatographic resolution and its results are easily interpretable using libraries of molecular fragmentation patterns [34].

#### *4.3. Proteome*

The term "proteome" encompasses all proteins in the oral cavity. Saliva contains more than 2000 proteins with a multitude of biological activities [35] and one quarter of salivary proteins are detectable in plasma. The greatest obstacle to salivary proteome analysis is its rapid degradation, which occurs just minutes after sample collection. For this reason, the majority of researchers combine the saliva with protease inhibitor cocktails (PIC) before storage and analysis, as suggested by Xiao et al. [36]. Proteomics takes advantage of NMR spectroscopy and gas and liquid chromatography–mass spectrometry (GC-MS and LC-MS), described above. In this research field, two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) and capillary electrophoresis with electrochemical detection are essential tools [37]. 2D-PAGE, which precedes the advent of 2D-difference gel electrophoresis (2D-DIGE), fractionates proteins on the basis of their isoelectric points in the first dimension and apparent molecular weight in the second [38]. An amphoteric carrier or an ampholyte is added to a gel and subjected to electrophoresis under a continuously regulated temperature. The acrylamide gel is placed in a glass tube and proteins are separated via an isoelectric gradient; it is easy to understand why this method is poorly accurate with multiple samples. 2D-DIGE dramatically improved 2D-PAGE thanks to the possibility of labeling each sample with distinct fluorescent dyes and then reading them using a laser scanner. Immunoassay is one of the most commonly used analytical techniques to detect the expression of an antibody or an antigen in a test sample. Enzyme-linked immunosorbent assay (ELISA) has been used for a variety of applications including diagnostic tools and quality controls [39]. The four basic setups are direct, indirect, sandwich, and competitive ELISAs. Direct ELISA is the

simplest format, requiring an antigen and an enzyme-conjugated antibody specific to the antigen [40]. The ELISA method is a sensitive and specific test that rapidly produces results and for these advantages has found a wide field of applications in clinical practice (e.g., in viral serology tests).

#### *4.4. Microbiome*

The study of microbiota has probably been the largest topic in scientific literature in recent years. In fact, next-generation sequencing has allowed the identification of thousands of phylotypes of microorganism throughout the entire human body, and research is ongoing. About 19,000 microorganisms have been identified in saliva [41]. Oral dysbiosis can lead to periodontal disease [42], caries [43], and some evidence exists supporting an association with cancer and systemic diseases [44,45]. Nowadays, molecular biology methods such as 16S ribosomal RNA (rRNA) gene sequencing, polymerase chain reaction (PCR), and other related PCR-based methods are very popular thanks to their high sensitivity and reproducibility. However, these techniques are no longer employed in routine diagnostics due to their costs. Alternative approaches include electromigration techniques (two-dimensional gel electrophoresis, capillary zone electrophoresis) and MS methods, such as matrix-assisted laser desorption ionization time-of-flight mode (MALDI-TOF MS). MALDI-TOF MS is a fast and accurate method based on the ionization of intact microorganism cells with short laser pulses and the subsequent acceleration of the particles in a vacuum by way of an electric field. Each microorganism has a specific spectrum profile [46].

Histopathology, in some cases with direct immunofluorescence, remains the gold standard for the diagnosis of oral disease. In fact, it is often necessary to perform a biopsy to confirm the diagnosis of bullous diseases (together with DIF) [47], Sjögren's syndrome [48], and for all lesions suspected for malignancy [49].

#### **5. Fields of Application**

In this article, we have summarized the latest findings on the use of saliva as a diagnostic tool in oral inflammatory diseases. In particular, we chose the most epidemiologically relevant conditions or where the oral cavity is a typical location of a systemic disease. In fact, mouth disorders can often precede the onset of systemic symptoms (e.g., in bullous pemphigus), and early diagnosis of oral disease can change the prognosis of these patients. In this scenario, the study of salivary biomarkers is a promising tool for early diagnosis and screening in susceptible populations (e.g., in smokers).

#### *5.1. Oral Lichen Planus*

Oral lichen planus (OLP) is one of the most common chronic inflammatory condition of the oral mucosa, with 0.5%–2% prevalence in adults and a slight predominance in women [50]. OLP affects oral mucosa symmetrically, with a predilection for oral mucosa. Clinically, it is possible to distinguish different aspects: reticular (the most prevalent form), erythematous, ulcerative or erosive, plaque-like, bullous, or papular [51,52]. The histopathology of OLP is typical, with a prominent lymphocyte infiltrate at the interface of epithelium, acanthosis, and degeneration of the basal cell layer [53]. Direct immunofluorescence (DIF) permits deposition of Immunoglobulin M as colloid bodies and C3 in granular and linear patterns in the basement membrane zone to be detected [54]. Although the exact pathogenesis of OLP is mostly unknown, it is believed that autoreactive T cells play a crucial role in the disease. Several risks/triggers factors have been described, such as stress, HCV and viral infections, and drugs [55]. OLP has been classified as a premalignant lesion for its risk of malignant transformation (0.04–1.74 per year) in squamous cell carcinoma (OSCC) [56]. Patients affected by OLP suffer from burning and itching sensations up to a severe pain in the erosive form; the disease has a huge negative impact on quality of life due to impairment in daily activities such as eating or oral hygiene [57]. Published articles focused on salivary biomarkers in OLP are quite recent and concern the diagnosis of OLP, but in particular the early detection of malignant transformation.

In 2018, Sineepat et al. enrolled five OLP patients and five healthy controls using a proteomic approach on saliva with two-dimensional gel electrophoresis followed by mass spectrometry. The authors detected three proteins that showed a potential role in OLP patients (cystatin SA, chain C of human complement component C3c, and chain B of fibrinogen fragment D) and tested with ELISA. All the analytical techniques confirmed with statistical significance that fibrinogen fragment D and complement component C3c were increased and cystatin SA was decreased in OLP patients compared with healthy subjects [58]. Fibrinogen fragments D and C3c play a central role in inflammation, whereas cystatin SA belongs to the cystatin superfamily, a group of cysteine protease inhibitors with antimicrobial activity. In fact, fibrinogen expression and C3 deposition are typical findings in OLP using IFD [54].

A different and more complex panel of proteins was reported by another study, published in 2017 [59]. The study was conducted on 10 patients, investigating with mass spectrometry 108 proteins differentially expressed in OLP subjects in comparison to healthy controls. The first finding was the absence of proteins essential to lubrication and viscoelasticity, supporting the xerostomia symptom frequently reported by patients. The authors interestingly tried to link protein expression in saliva with histological findings in OLP, discussing the known functions of each peptide. In particular, S100A8 and S100A9 (also called MRP8 and MRP14) are calcium- and zinc-binding proteins with a role in inflammation and cytokine production via IL-17. S100A8 can also induce apoptosis via attraction to skin of CD8<sup>+</sup> cells and natural killer (NK) cells [60]. Another player in the scene of T-cell proliferation and differentiation is AZGP1 (zinc-alpha-2-glycoprotein), which is an adipocytokine [61]. The study also confirmed the crucial role of oxidative stress in OLP; reactive oxygen species (ROS) induce apoptosis and dysfunction in keratinocytes and, moreover, ROS can be further produced from TCD4<sup>+</sup> lymphocytes infiltrating in OLP in a vicious circle.

Oxidative stress in OLP was previously discussed in 2016 in a case–control study enrolling 62 patients and 30 healthy individuals [62]. The authors demonstrated significant differences between patients and the control group concerning the average concentration of total antioxidant capacity (TAC, determined using the Benzie and Strain method [26]), glutathione (GSH, measured spectrophotometrically), and thiobarbituric-acid-reactive substances (TBARS, determined using the Aust method), which are a product of lipid peroxidation [63]. In patients suffering from OLP, as expected, TAC and GSH had lower values, while TBARS was higher than in healthy controls. More interestingly, patients with an erosive form of lichen had more marked values, demonstrating severe oxidative stress and a great concordance with clinical features. These findings could support the oral or topical use of antioxidants [64].

Many authors have suggested salivary cortisol as a biomarker in OLP [65–70]. Cortisol is considered a biological marker of stress and anxiety, the variation of which can alter cytokine profiles [71]. OLP has a double connection with stress: anxiety and stressful events are considered a trigger for OLP onset but, at the same time, oral lichen itself represent a source of stress for patients. In this intricate scenario, the evaluation of salivary cortisol seems to mimic the ancestral question "Which came first, the chicken or the egg?" In fact, data from the literature are controversial, and cortisol is probably not suitable as a biomarker in OLP. As previously discussed, OLP is a T-cell-driven disease; however, it is still unclear if the inflammation is due to Th1 or Th2 expression. In fact, in OLP, there are numerous cytokines expressed both from recruited lymphocytes and from affected keratinocytes, in a mechanism of self-amplification [72]. The evaluation of specific interleukin in saliva is certainly a good trace to detect biomarkers in OLP and, moreover, to design tailored therapies. Nowadays, more consistent results concern IL-6 and IL-8. Interleukin-6 is involved in B- and T-cell differentiation and is able to inactivate p53 with tumor progression of some cancers [73]. Mozaffari et al. revealed in a meta-analysis that IL-6 levels in saliva and serum of OLP patients were significantly higher than in healthy controls, with higher values in saliva than in serum [74]. For this reason, saliva seemed to be more useful than serum for the detection of IL-6. Interleukin-8 is an important mediator of host response to injury and inflammation; it can activate neutrophils, basophils, and T cells [75]. The group of Mozaffari conducted

a meta-analysis on this topic [76]. The most interesting finding was that IL-8 plays a key role in the transformation from reticular to erosive form of lichen, probably due to the loss of efficacy in the repairing mechanisms of keratinocytes [77]. IL-8 also revealed a potential application in therapeutic monitoring, as demonstrated via its decrease in saliva after dexamethasone administration [78].

#### *5.2. Oral Cancer*

Oral cancer is the sixth most common cancer worldwide [79] with a higher incidence in India, because of the chewing of areca nut/betel quid [80]. The mortality rate after 5 years from diagnosis is still 50% [79]. Well-known risk factors include tobacco consumption, alcohol abuse, and human papilloma virus infections [81]. The onset of an oral cancer is frequently asymptomatic, but most oral carcinomas develop from premalignant conditions such as leukoplakia and oral lichen planus [82,83]. Nowadays, the gold standard for diagnosis is tissue biopsy, an invasive technique that requires specific training and creates public health costs [84]. It is therefore easy to understand the need for an early detection method for pre-cancer and cancer by validating salivary biomarkers. Above, we discussed the great diagnostic potential of miRNA both in saliva and serum. In 2018, a well-designed study enrolled 30 patients with OLP, 15 patients with OSCC and 15 healthy donors [85]. Saliva samples were analyzed by quantitative RT-PCR for miR-21, miR-125a, miR31, and miR200a. Results showed that miRNA-21 and -125a were, respectively, higher and lower in OSCC patients and in OLP with dysplasia compared to healthy controls with statistical significance. miR-21 has been widely studied in oral, head, and neck cancer and has been postulated that it might have a role in inhibition of tumor suppression and apoptosis [86]. In contrast, miR-125a may act as a tumor suppressor, downregulating target oncogens [87]. Based on these data, the authors suggested a negative prognostic role of decreased salivary miR-125a levels in association with increased salivary miR-21 levels in OLP patients. Ishikawa et al. recently suggested a metabolomics approach to distinguish OLP from OSCC [88]; the authors detected higher levels of 12 salivary metabolites in OSCC patients compared with OLP patients. More specifically, the combination of indole-3-acetate and ethanolamine phosphate showed the best statistical accuracy. The aim of Mikkonen's research was to investigate the potential of nuclear magnetic resonance (NMR) spectroscopy for detecting the salivary metabolic changes associated with head and neck squamous cell carcinoma (HNSCC). The Authors found two metabolites, fucose and 1,2-propanediol, to be significantly upregulated, whereas proline was significantly downregulated in patients affected by HNSCC. The combination of four salivary metabolites (fucose, glycine, methanol, and proline) together provided maximum discrimination among HNSCC patients and healthy controls [31]. The role of fucosylation of glycoproteins in the development of cancer has been studied in recent years [89]. Ample evidence exists to prove that in normal tissues, fucosylation levels are relatively low, but this rapidly increases during carcinogenesis [90]. Aberrant glycosylation in cancer development is also an investigation area in oral diseases; in particular, researchers have focused attention on sialic acid (N-acetyl neuraminic acid), which is an important terminal sugar in cell membrane glycoproteins and glycolipids. Previous studies have shown elevated levels of salivary sialic acid in various carcinomas, including oral pre-cancer and OC [91–93].

The fascinating study of the microbiome has a wide field of application in the oral cavity. An extensive work has just been published regarding the alterations of salivary microbial community in oropharyngeal and hypopharyngeal carcinoma patients. In fact, the microbiome is considered a potential modulator of cancer metabolism [94]. The authors found 13 phylotypes of microorganism as potential diagnostic biomarkers in oral cancer. The role of the microbiome in malignant change in the oral cavity is still controversial because of the lack of large cohort studies. Healy et al. considered the implication of risk factors such as smoking or alcohol consumption in promoting epithelial dysplasia and production of carcinogenic agents [95]. Acetaldehyde (ACH) and N-nitrosamine compounds are potential genotoxic agents that are increased in the saliva of smokers; these compounds can be produced in vitro by microbial cultures [96,97]. In vitro studies have demonstrated the leading role of *Neisseria* species and *Candida* species in ACH production [98,99]. However, one study revealed the

reduction of *Neisseria* species in the oral cavity of smokers, with a theoretical improvement of ACH levels [100]. Current theories hypothesize that the presence of these organisms could accelerate the progression of dysplasia towards OSCC in association with predisposing factors such as diet, age, or smoking/alcohol consumption habits in a multifactorial vision.

#### *5.3. Blistering Diseases*

Bullous pemphigoid (BP) and pemphigus vulgaris (PV) are acquired bullous diseases affecting the mucosa and/or skin. In both diseases, autoantibodies react with adhesion cell mechanisms or with the basement layer, resulting in blistering. Blisters are intraepithelial/intraepidermal in PV, whereas in BP they are subepithelial/subepidermal [101]. The diagnosis is first clinical, then confirmed with histopathology and direct immunofluorescence (IFD). In BP, bullae involving the skin and oral lesions are rare; in contrast, PV frequently begins with oral blistering or oral lesions following cutaneous involvement. IFD reveals IgG and C3 (BP180) deposition on the basement membrane in BP, while in PV it shows intercellular IgG antibody deposition to desmoglein (Dsg) 1 and/or desmoglein 3, which are trans-membrane desmosomal proteins [102]. In recent years, the use of ELISA to detect autoantibodies in the serum of BP and PV patients has entered clinical practice for diagnosis and therapeutic monitoring [101]. Starting from this technique, some authors have proposed the use of saliva as substrate for the research of BP180 and Dsg1 and 3. In 2006, Andreadis et al. first applied ELISA in both the serum and saliva of PV and BP patients, finding a great concordance in serum and saliva levels of Dsg1 and 3, while the BP180 determination on saliva failed [103]. Similar results emerged from Ali's study [104] on Dsg1 and 3. The potential of salivary testing in PV prognosis and mucosal severity has been investigated in two studies. Hallaji et al. included 50 patients with histologically confirmed PV and performed ELISA for Dsg1 and 3 on serum and saliva samples [105]. There was statistically significant concordance between serum and salivary levels of Dsg; more interestingly, there was a significant relationship between salivary anti-Dsg1 antibody and mucosal severity. The authors explained these data with the loss of integrity in mucosa and the largest transition of antibodies in saliva. The study of De et al. perfectly reproduced this finding and the authors perfectly agreed with the explanation concerning higher Dsg1 levels in severe disease [106]. In contrast to the previously discussed research, one Italian study was designed to assess the use of a BIOCHIP approach compared with ELISA in PV [107]. In fact, the authors considered saliva an unsuitable substrate for autoantibody detection because of the discordance between techniques found when using saliva samples.

#### *5.4. Sjögren's Syndrome*

Sjögren's syndrome (SS) is a systemic autoimmune disease characterized by the inflammation and consecutive destruction of exocrine glands, as well as salivary and lacrimal glands, with the occurrence of a lymphoepithelial sialadenitis [108]. The majority of patients are women of menopausal age; oral manifestations are frequently present at the onset of disease, but some patients develop a systemic disease with the involvement of joints, the gastrointestinal tract, the central nervous system, and with an increased risk of lymphoma [109]. Patients suffering from SS typically complain about xerostomia and its impact on their quality of life [110]. Current research on salivary biomarkers in SS is pursuing a non-invasive diagnostic test, a therapeutic monitoring marker, and, moreover, an early detection of lymphoma onset. One of the current diagnostic approaches is the detection of anti-Ro/SSA and/or anti-La/SSB in serum; studies from different groups have demonstrated the presence of these autoantibodies in the saliva of SS patients [111,112]. The determination of salivary autoantibodies seemed to be effective in discriminating SS patients from patients affected by systemic lupus erythematosus (SLE) [113]. A few studies have investigated cytokine profiles in SS saliva; data from these studies showed significantly higher levels of Th1, Th2, and Th17, in accordance with serum findings [114,115]. The proteomic approach in SS comprises proteins, enzymes, calcium-binding proteins, and immune-related molecules. Summarizing, data from the literature report high levels of inflammatory-phase proteins in saliva that can provide a great indication of gland status [116]. Lee et al. recently published the results of determination of soluble sialic-acid-binding immunoglobulin-like lectin (siglec)-5 in saliva and sera by ELISA [117]. The level of salivary siglec-5 was significantly higher in the saliva from SS patients, which reflects the severity of hyposalivation. Several novel miRNAs have been described in SS [118]. Pauley et al. demonstrated that the expression of miR-146a was significantly increased in SS patients [119]. In Alevizos' research, another two miRNAs, miR-768-3p and miR-574, were associated with minor salivary gland inflammation in 15 patients with SS [120]. The pathogenesis of autoimmune diseases is a very complex interaction of many factors; epigenetic modifications are now considered crucial to the control of gene expression associated with these diseases [121,122]. Thabet et al. proposed that the dysfunction of salivary gland epithelial cells in SS might be partially linked to epigenetic modifications. Their analysis showed that blood global DNA methylation was reduced in SS patients and the expression of the gene DNMT1, which encodes DNA methyltransferase 1, was decreased compared to healthy controls. In contrast, the expression of the gene Gadd45a, which encodes the growth arrest and DNA-damage-inducible protein GADD45 alpha (GADD45a), was increased [123]. Probably the most interesting field in saliva and SS is the early diagnosis and prevention of MALT-type lymphoma [124]. The neoplasm has an insidious onset, almost asymptomatic, with a fast progression and dissemination. Cui et al. described a triad of markers (anti-cofillin-1, anti-alpha-enolase, and anti-Rho GDP-dissociation inhibitor 2) overexpressed in patients with SS who developed MALT lymphoma compared with SS patients and healthy individuals [125]. Sharma et al. recently examined the role of the microbiome in SS compared to healthy controls [126]. The analysis, performed with DNA isolation and 16rRNA sequencing, revealed four genera (Bifidobacterium, Dialister, Lactobacillus, and Leptotrichia) that were different between the two groups. The results were consistent with previous studies, revealing a role of Actinobacteria and Firmicutes phila [127,128]. More interestingly, Sharma et al. identified a difference in alpha diversity in patients treated with steroids, suggesting the potential role of microbiome analysis in therapeutic response.

#### *5.5. Psoriasis*

Psoriasis is now classified as an immune-mediated inflammatory disease (IMID) of the skin. It is being recognized that patients with psoriasis are at higher risk of developing systemic co-morbidities, e.g., metabolic syndrome and cardiovascular diseases [129,130].

Oral involvement in course of psoriasis is still debated. Recently, it has been hypothesized that gingivitis and periodontitis share the same underlying inflammatory pathogenic process as psoriasis. Thus, in our previous study, psoriatic patients were investigated for oral mucosa lesion prevalence as well as gum disease. Results displayed an increased association between gingivitis/periodontitis and psoriasis, which may suggest common underlying pathogenic risk factors [131].

Furthermore, salivary secretions, collected from patients with active psoriasis and healthy control subjects, were investigated for expression of interleukin (IL)-1b, IL-6, transforming growth factor (TGF)-β1, IL-8, tumor necrosis factor (TNF)-α, interferon (IFN)-χ, IL-17A, IL-4, IL-10, monocyte chemoattractant protein (MCP)-1, microphage inflammatory protein (MIP)-1a, and MIP-1b using a Multi-Analyte ELISArray Kit (Qiagen, Venlo, the Netherlands). Patients with active psoriasis had significantly higher salivary IL1β, TNF-α, TGF-β, and MCP-1 levels than healthy controls [132].

Thus, saliva can be a valid non-invasive tool for monitoring inflammation in psoriasis [133].

#### **6. Conclusions**

In the era of precision medicine, salivaomics approaches seem to be a promising field of research. Despite encouraging results reported in this review, there is a large variability in study designs, protocols, sampling collections, and techniques. Moreover, the study of new molecules with new technologies requires a well-established range of values without random decisions. Future studies should standardize accurate methodologies in order to validate new salivary biomarkers in clinical practice.

**Author Contributions:** E.M. was responsible for the literature revision and drafting; A.C. was responsible for final drafting and revision; F.D. contributed to article analysis; A.O. gave her final revision. All authors have read and agree to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest to declare.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Journal of Clinical Medicine* Editorial Office E-mail: jcm@mdpi.com www.mdpi.com/journal/jcm

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18

www.mdpi.com ISBN 978-3-0365-0777-4